A4F
v16.0.0•

Flagship Model Expansion & Responses API Launch

Sree
SreeAuthor

This milestone update delivers one of our most significant platform upgrades, introducing the latest flagship models from leading AI research labs, expanding capabilities with advanced audio and image processing, and launching a purpose-built API for agentic workflows.

GPT-5 Family - Next-Generation Flagship Models

8

GPT-4o Family - Enhanced Multimodal Models

7

GPT-4.1 Family - Advanced Vision & Function Calling

6

Reasoning-Focused O-Series Models

7

GPT-3.5 Family - Cost-Effective Chat Models

6

Specialized Models

3

Additional Platform Updates

5

Provider 1 Updates

7
  • Catalog Refocus: Complete restructuring to exclusively feature premium, enterprise-grade models from tier-1 AI labs. Retired open-source and community-tuned models for curated, production-ready catalog.
  • Mistral AI Models: Added mistral-large-latest, codestral-latest, and pixtral-large-latest with multimodal capabilities.
  • GPT-4.1 Family: Integrated complete GPT-4.1 series with improved instruction following and enhanced reasoning across various model sizes.
  • OpenAI Open-Source: Added gpt-oss-120b, OpenAI's 120B parameter open-source model for research and development.
  • DeepSeek Models: Integrated deepseek-v3.2-exp experimental variant and deepseek-r1-0528 reasoning-optimized model.
  • Zhipu AI Integration: Added glm-4.6 with enhanced multilingual capabilities and improved reasoning performance.
  • Production Validation: All models validated for production workloads with established SLAs, comprehensive documentation, and dedicated support.

Provider 3 Updates

5
  • Model Graduation: Promoted qwen-3-max-preview to production-ready qwen-3-max with full SLA guarantees and enterprise support.
  • Legacy Retirement: Removed legacy models to streamline catalog:
  • Enhanced Reliability: Improved uptime guarantees, reduced latency, and enhanced error handling for production deployments.
  • Multi-Modal Support: Comprehensive support for text generation, image creation, image editing, and multimodal tasks.
  • Performance Optimization: Improved request routing, enhanced caching strategies, and optimized model loading for reduced cold-start times.

Provider 5 Updates

6
  • Strategic Refocus: Reimagined as dedicated channel for direct OpenAI and Midjourney integrations with comprehensive, version-pinned catalogs.
  • Streamlined Scope: Removed all xAI (Grok) and Anthropic (Claude) models to focus exclusively on OpenAI and Midjourney offerings.
  • Complete GPT Coverage: Full deployment including gpt-3.5-turbo, gpt-4-turbo, gpt-4o, and entire gpt-5 family with version-pinned variants.
  • Image Generation: Added dall-e-2 and dall-e-3 for text-to-image generation.
  • Embedding Support: Deployed text-embedding-ada-002 for semantic search, document similarity, and RAG applications.
  • Direct API Benefits: Lowest latency, highest reliability, and most comprehensive feature support through direct API integration.

Provider 7 Updates

1
  • Latest Claude Variant: Added claude-haiku-4-5-20251001 with improved speed, enhanced reasoning, and optimized cost-efficiency, available exclusively for Ultra plan subscribers.

Plan Availability Changes

5

Improvements

6
  • Enhanced Model Routing: Intelligent request routing across providers based on availability, load, and performance for optimal response times.
  • Improved Error Handling: Comprehensive error handling with detailed messages, automatic retry logic, and graceful degradation.
  • Performance Optimization: Provider-level optimizations including connection pooling, request batching, and intelligent caching for reduced latency.
  • Dashboard Redesign: Completely redesigned dashboard with compact, professional layout featuring improved information density, optimized spacing, and enhanced visual hierarchy for better user experience.
  • Documentation Updates: Complete documentation refresh covering all new models, capabilities, and API endpoints with examples and migration guides.
  • Responses API Documentation: Added comprehensive documentation page for the new /v1/responses endpoint with detailed request/response examples, parameter descriptions, and usage guidelines.